On 18 Dec 2016, at 17:37, Ed Schouten <e...@nuxi.nl> wrote: > > 2016-12-18 16:21 GMT+01:00 Dimitry Andric <d...@freebsd.org>: >> Modified: head/sys/dev/mlx4/mlx4_ib/mlx4_ib_sysfs.c >> ============================================================================== >> --- head/sys/dev/mlx4/mlx4_ib/mlx4_ib_sysfs.c Sun Dec 18 15:10:08 2016 >> (r310231) >> +++ head/sys/dev/mlx4/mlx4_ib/mlx4_ib_sysfs.c Sun Dec 18 15:21:38 2016 >> (r310232) >> @@ -79,7 +79,7 @@ static ssize_t store_admin_alias_guid(st >> container_of(attr, struct mlx4_ib_iov_sysfs_attr, dentry); >> struct mlx4_ib_iov_port *port = mlx4_ib_iov_dentry->ctx; >> struct mlx4_ib_dev *mdev = port->dev; >> - u64 sysadmin_ag_val; >> + unsigned long long sysadmin_ag_val; >> >> record_num = mlx4_ib_iov_dentry->entry_num / 8; >> guid_index_in_rec = mlx4_ib_iov_dentry->entry_num % 8; > > Looking at the code, the intent is that an actual 64-bit integer is > parsed; not an unsigned long long. What's wrong with using > inttypes.h's SCNx64?
That is how I originally submitted it in the review, though with <machine/_inttypes.h>, as it is most often used in kernel sources. However, see Hans's comment in the review here: https://reviews.freebsd.org/D8831#182872 I don't mind either way. -Dimitry
signature.asc
Description: Message signed with OpenPGP using GPGMail